Systems that detect fraud in Medicare and Medicaid to be integrated
The federal office that combats fraud in Medicare and Medicaid has begun to integrate its two predictive data models, said David Nelson of the Centers for Medicare and Medicaid Services.
Nelson heads the the data analytics and control group for CMS's Center for Program Integrity. He spoke Sept. 18 at the Predictive Analytics World-Government conference in Washington, D.C.
CPI uses one automated system to flag those healthcare providers most likely to commit fraud because of past criminal and financial issues, among other factors. Another system rates fee-for-service claims for their risk of fraud.
The systems were intended to work together, Nelson said. The risk score that the first system gives a provider could be factored into the second system's rating of a specific claim's risk. "That, we're just starting to do right now," Nelson said, though CPI hasn't fully integrated the systems yet.
The system that looks at claims uses dozens of different models to find fraud, and Nelson said CPI plans to more than double that figure by the end of the year.
He also gave some examples of how the system has found fraudulent claims so far. In one case, the provider and the patient were in different states, and the service involved was one that usually entailed an extensive relationship between the provider and patient.
Another scam caught the system's attention because patients were getting a certain service at eight times the national average.
CPI refers the cases it finds to the Health and Human Services Department office of inspector general and has trained OIG officials to use the predictive systems.
In July, CMS also launched a new command center where CPI, OIG, law enforcement, the FBI, policy specialists and contractors work together with CPI's predictive systems.
By the end of September, CPI will publish its first report to Congress on its predictive data efforts. Nelson said the center is currently putting the final touches on the report.